In a recent article, GigaOm’s Stacey Higginbotham writes about SDN (and more specifically, OpenFlow) and commoditization. She asserts the following:
“If 2012 was the year that software-defined networking sold out, then 2013 should be the year that the big players in the industry recognize that their efforts to neutralize the threat of OpenFlow and the coming commoditization of networking hardware are doomed to failure.”
Stacey’s overall point – that SDN is going to happen – is right, but I don’t think that commoditization is going to be the driving force. If you follow where the money is going, you can see that a lot of people are betting on SDN solutions emerging. For the first time in a couple of decades, we are seeing substantial enough opportunities (fueled no doubt by recent industry consolidation) to lure away long-time fixtures at the larger players.
But if money is flowing into companies who want a piece of the action, and talent is following, what will they base their business models on? I cannot imagine that building a business around the commoditization of hardware is the best way to get great returns. There has to be something more behind the push. The real question is: what is that something?
I think it’s pretty simple, really: the switching market at large is the most underserved market around. There is no market in any industry, high tech or other, where customers are so acutely aware of their vendor margins. That’s not to say that there is no other market that is price sensitive, but the hyperawareness of vendor margins in the networking space is fairly unique. In the consumer space, we don’t see millions of iPhone users begrudging Apple their profit, but in the switching space, we care. Why is that?
Let’s be honest for a minute. The space is dominated by a single vendor. The lack of any kind of sustained competition has led to an approach to solving problems that is largely incremental. Start with some basic solution, run into a problem, add a knob. Or a protocol. Or whatever. Over time, the products haven’t moved forward by the same leaps and bounds that other areas (like compute and storage) have. The result? Users are left rubbing their heads, knowing they need more but unable to do anything about it. Once you accept that you aren’t going to get more, the best you can hope for is to get the same thing a little bit cheaper.
With all of these bets going after a big market, what will separate the sharks from the chum is how well they understand what the market really needs. That is a topic for another post, but I’ll leave you this little nugget: the answer is not more protocols or a slightly cheaper version of the same thing we have had for years.